Metix AI · Discovery Loop Talent BriefMetix AI · Discovery Loop Talent Brief

Jeff Dean 离开 Google,揭示 AI 下一站人才缺口Jeff Dean leaves Google, revealing AI’s next talent gap
AI 竞争正走向模型、系统与科学的闭环AI competition is moving toward a closed loop of models, systems, and science

Jeff Dean、Sanjay Ghemawat、Quoc V. Le 和 Oriol Vinyals 的公开履历指向四类能力:研究系统化、底层基础设施、模型自动化、智能体与评测。Metix AI 的人才数据覆盖这四类能力的完整梯度——从人才池最厚的基础设施,到 Model Serving、AutoML / NAS、Deep RL、Seq2seq / NLP 这类更专精的方向,都能精确定位。Jeff Dean, Sanjay Ghemawat, Quoc V. Le, and Oriol Vinyals' public track records point to four capabilities: research systemization, low-level infrastructure, model automation, and agents/evaluation. Metix AI's talent data spans the full depth of these four capabilities — from infrastructure, the deepest pool, to more specialized tracks like Model Serving, AutoML/NAS, Deep RL, and Seq2seq/NLP, all precisely locatable.

Focus Discovery Loop talent stackSupply depth 594 verified · 12 rolesAs of 2026-08-08
Founder stack

01四位创始人的公开履历,指向四类可招聘能力The four founders’ public track records point to four recruitable capabilities

Jeff Dean、Sanjay Ghemawat、Quoc V. Le、Oriol Vinyals 的公开履历覆盖研究系统、底层基础设施、模型自动化、智能体评测。Discovery Loop 要形成从研究设想到实验反馈的循环,四类能力需要同时进入组织。Jeff Dean, Sanjay Ghemawat, Quoc V. Le, and Oriol Vinyals bring public track records across research systems, low-level infrastructure, model automation, and agent evaluation. A Discovery Loop-style organization needs these four capabilities inside the same operating system from research idea to experimental feedback.

4
创始人对应四类核心能力Four founders, four core capabilities
覆盖 12 个可招聘岗位Spanning 12 hireable roles
594
项目血统覆盖的可核实人才Verified talent across the project lineage
397 精准匹配397 precisely matched
264
同类公司招聘职位Postings at peer companies
21 家同类公司 · 2026-08-0821 peer companies · 2026-08-08
3,612
同类公司人才记录规模Peer-company talent-record pool
前四家公司占 50.2%Top four companies hold 50.2%
01四类能力已经对应真实、可核实的候选人The four capabilities already map to a real, verifiable talent pool

594 人分布在 12 个具体岗位上,覆盖美国、英国、加拿大等六个国家——创始人的履历框架,已经能落到今天真实的招聘市场上。Those 594 people sit across 12 concrete roles in six countries, including the U.S., U.K., and Canada — the founders' capability framework already holds up in today's real hiring market.

02人才供给集中在底层基础设施,模型自动化和智能体最稀薄Supply concentrates in low-level infrastructure; model automation and agents are thinnest

底层基础设施相关的 4 个岗位合计 276 人,模型自动化和智能体两组加起来只有 145 人——越靠近训练基础设施,可招聘的人越多;越靠近前沿研究方向,人才池越窄。The 4 low-level infrastructure roles hold 276 people combined, while model automation and agents together hold only 145 — the closer a role sits to training infrastructure, the deeper the hireable pool; the closer to frontier research, the narrower it gets.

03同类公司的招聘重心落在基础设施和领域科学Peer-company hiring centers on infrastructure and domain science

264 个招聘职位中,Infrastructure/Platform(96)和 Domain Science(85)合计过半,资历集中在 Mid-Senior level(183)——需求方向与四位创始人代表的能力组合高度吻合。Of 264 postings, Infrastructure/Platform (96) and Domain Science (85) together account for more than half, with seniority concentrated at Mid-Senior level (183) — demand closely mirrors the capability mix the four founders represent.

04人才供给集中在少数几家公司Talent supply concentrates in a handful of companies

21 家同类公司合计 3,612 条人才记录,Recursion、Lila Sciences、Isomorphic Labs、Generate:Biomedicines 四家就占 50.2%——优先从这四家切入,比撒向全部 21 家公司更高效。21 peer companies hold 3,612 talent records combined; Recursion, Lila Sciences, Isomorphic Labs, and Generate:Biomedicines alone account for 50.2% — starting there is more efficient than spreading across all 21 companies.

"自动化发现,加速全球科学与工程"——Discovery Loop"Automating discovery to accelerate science and engineering for the world" — Discovery Loop

四类能力形成从左到右的链路:研究系统化承接模型与平台,底层基础设施支撑数据和调度,模型自动化扩大搜索空间,智能体与评测把结果拉回可验证任务。The four capabilities form a left-to-right chain: research systemization connects models and platforms, low-level infrastructure supports data and scheduling, model automation expands the search space, and agents plus evaluation bring results back into verifiable tasks.

01研究系统化Research systemizationJeff Dean

把研究想法变成全球级 AI 系统,贯通训练与部署全流程。Turns research ideas into global-scale AI systems, spanning training through deployment.

Google BrainDistBeliefTensorFlowPathwaysTPUPaLM/GeminiML at ScaleML Platforms
02底层基础设施Low-level infrastructureSanjay Ghemawat

解决数据流、存储、调度、可靠性和性能。Covers dataflow, storage, scheduling, reliability, and performance.

MapReduceBigtableSpannerPathwaysRPC systemsPerf ToolsData/Storage
03模型自动化Model automationQuoc V. Le

让模型结构、训练方法和推理能力可以规模化搜索和迁移。Scales search and transfer across model architecture, training methods, and reasoning.

Seq2seqNMTNASAutoMLFLANEfficientNetGLaMAlphaGeom
04智能体与评测Agents and evaluationOriol Vinyals

把序列建模、强化学习、多模态和代码推理落到可验证任务。Brings sequence modeling and reinforcement learning into verifiable, multimodal tasks.

Seq2seqAlphaStarAlphaCodedistillationTensorFlowDeep RLMultimodalEval/Bench
Talent lineage

02项目血统还原出 12 个可招聘的岗位画像Project lineage resolves into 12 hireable role profiles

项目会过时,能力不会。MapReduce 和 RPC 系统是上一个技术周期的产物,但支撑它们的分布式数据处理能力和跨节点通信能力,如今换了一层包装,以"大规模训练数据流水线工程"和"GPU 集群网络通信工程"的新形态,重新出现在今天的招聘市场上。四位创始人的能力谱系,收敛成四组、12 个可招聘的岗位。Projects go out of date, capability doesn't. MapReduce and Google's RPC systems belong to the last technology cycle, but the distributed data processing and cross-node communication capability behind them has been repackaged — reappearing today as large-scale training data pipelines and GPU-cluster networking. The four founders' capability lineage resolves into four groups of 12 hireable roles.

创始人—项目—岗位关系图谱Founders, Projects, and Roles at a Glance

模型自动化方向记录了 11 个知名项目,是四组里最多的,但项目大多诞生于最近两三年(Gemini、AlphaGeometry 都是 2023 年之后的工作),这个方向的从业规模还在早期,眼下能招到的 80 人已经站在最前排;底层基础设施的项目血统可以追溯到 2003 年的 GFS,二十年的积累让 276 人的人才池自然更深。Model automation traces back to 11 landmark projects, the most of any group — but most were built in just the last two or three years (Gemini and AlphaGeometry are both post-2023 work), so the field's talent base is still early-stage; the 80 people hireable today are already at the front of it. Low-level infrastructure's project lineage reaches back to GFS in 2003 — two decades of accumulation is what makes its 276-person pool run deep.

Drag to rearrange · Hover to highlight · Data: Metix AI
Jeff Dean Sanjay Ghemawat Quoc V. Le Oriol Vinyals Infrastructure Training Systems Model Automation Agents & Eval
底层基础设施Low-level infrastructure4 岗位4 roles

存储与网络最深,性能工程最窄Storage and networking deepest, performance narrowest

存储与数据库系统(109)和网络通信(103)撑起这组一半以上的人才池,但性能工程只有 19 人——同一条底层基础设施血统里,越靠近"调优"这类隐性技能,供给就越稀。Storage/database systems (109) and networking (103) carry more than half of this group's pool, but performance engineering has only 19 — within the same infrastructure lineage, supply thins out fastest for tacit tuning skills.

Distributed storage与数据库系统Distributed storage & database
10984强/25弱84 strong/25 weak
分布式训练网络通信Training networking (NCCL/RDMA)
10358强/45弱58 strong/45 weak
AI训练数据流水线AI training data pipeline
4528强/17弱28 strong/17 weak
AI训练/推理性能工程Training/inference performance
1911强/8弱11 strong/8 weak
数据来源: Metix AIData source: Metix AI
模型自动化Model automation4 岗位4 roles

越具体的模型研究方向,匹配越精确The more specific the direction, the cleaner the match

大语言模型预训练(87.5% 强匹配)和模型效率研究(100% 强匹配)人数都不多,但匹配精度是全部 12 个岗位里最高的两个——这类头衔本身已经足够具体,不需要靠宽泛推测。LLM pretraining (87.5% strong) and model efficiency (100% strong) are both small, but they're the two cleanest matches of all 12 roles — the titles themselves are specific enough that little guesswork is needed.

大语言模型研发(预训练)LLM pretraining research
2421强/3弱21 strong/3 weak
模型效率研究员Model efficiency research
99强/0弱9 strong/0 weak
指令微调/对齐研究员Instruction tuning / alignment
3118强/13弱18 strong/13 weak
AI数学/几何推理研究员AI math / geometry reasoning
168强/8弱8 strong/8 weak
数据来源: Metix AIData source: Metix AI
规模化训练系统Large-scale training systems2 岗位2 roles

TPU 是 12 个岗位里最大也最稳的池子TPU is the largest, steadiest pool of all 12

AI 加速器系统工程师(TPU)总数 117 人、强匹配占 82%,是全部 12 个岗位里规模最大、精度也最高的一个——TPU 作为独立岗位在招聘市场上有清晰的行业共识,这正是它更容易被精确统计到的原因。AI Accelerator Systems (TPU) totals 117 people at 82% strong-match — the largest and most precisely countable of all 12 roles, because TPU already reads as a distinct, well-understood job title across the industry.

AI加速器系统工程师(TPU)AI accelerator systems (TPU)
11796强/21弱96 strong/21 weak
大规模分布式训练系统Large-scale training systems
5642强/14弱42 strong/14 weak
数据来源: Metix AIData source: Metix AI
智能体与评测Agents & evaluation2 岗位2 roles

强化学习标签最模糊,代码推理最稀缺RL is the fuzziest label, code reasoning the scarcest

强化学习(推理后训练)研究员总数 60 人但强匹配只占 32%,是全部 12 个岗位里精度最低的一个——RL 相关头衔太容易被泛化引用;代码推理/生成研究员则是另一个极端,全球范围内只找到 5 人,是 12 个岗位里最稀缺的方向。Reinforcement learning (post-training) totals 60 but only 32% strong-match, the least precise of all 12 — RL-adjacent titles get invoked too loosely. Code reasoning/generation is the opposite extreme: only 5 people found worldwide, the scarcest of the 12.

强化学习(推理后训练)RL (post-training)
6019强/41弱19 strong/41 weak
代码推理/生成研究员Code reasoning / generation
53强/2弱3 strong/2 weak
数据来源: Metix AIData source: Metix AI

所有岗位均采用同一套检索方法和判定基准迭代核实得出,确保组间可比。数据来源: Metix AI(Mira 人才检索),统计范围为美国、英国、加拿大、法国、荷兰和日本的当前在职人才,截至 2026-08。Every role was verified with the same search method and matching bar, applied iteratively, so the groups stay comparable. Data source: Metix AI (Mira talent search); scope covers currently working people in the United States, United Kingdom, Canada, France, Netherlands, and Japan, as of 2026-08.

四个代表性岗位的强匹配样本Strong-match samples from four representative roles

按四个代表性岗位各挑 6 位强匹配人才:TPU 对应 Jeff Dean 的标志性项目;存储与数据库对应 Jeff Dean 与 Sanjay Ghemawat 共同署名的 Bigtable / Spanner;NCCL/RDMA 正是上文"项目血统"举的例子——RPC 系统演变为 GPU 集群网络通信;大语言模型研发(预训练)是 Jeff Dean、Oriol Vinyals、Quoc V. Le 三位创始人能力交汇的岗位。6 strong-match people from each of four representative roles: TPU maps to Jeff Dean's signature project; storage and databases map to Bigtable / Spanner, co-authored by Jeff Dean and Sanjay Ghemawat; NCCL/RDMA is the literal example used in the project-lineage narrative above — RPC systems evolving into GPU-cluster networking; LLM pretraining research is where Jeff Dean, Oriol Vinyals, and Quoc V. Le all converge.

AI加速器系统工程师(TPU)AI Accelerator Systems (TPU)

T●● S●● TG-01TPU
Senior Software Engineer - TPU Compiler at Google
美国United StatesTPU CompilerXLA / compiler
TPU Compiler团队的核心编译器研发,直接对应TPU项目血统。Core compiler work on the TPU Compiler team — a direct match to the TPU project lineage.
J●● S●● TG-02TPU
Senior Staff Software Engineer, XLA TPU Compiler
美国United StatesXLATPU Compiler
XLA TPU编译器方向的资深主任级工程师信号。Senior staff-level engineering signal on the XLA TPU compiler.
K●● N●● TG-03TPU
SW Engineer, Cloud TPU — Google, AMD, Ex-Meta, Ex-Intel
美国United StatesCloud TPUGoogle, AMD, Meta, Intel
跨Google、AMD、Meta、Intel的云端TPU与芯片工程经历。Cloud TPU and chip engineering experience spanning Google, AMD, Meta, and Intel.
R●● W●● TG-04TPU
RTL Design Engineer for AWS Trainium
美国United StatesRTL designAWS Trainium
AWS Trainium的RTL芯片设计,属于同类AI加速器硬件方向。RTL chip design for AWS Trainium — the same AI-accelerator hardware lineage.
S●● K●● TG-05TPU
ASIC Engineer, Infra Silicon at Meta, Ex Google TPU Lead, PhD
美国United StatesASICEx Google TPU Lead
前Google TPU团队负责人,现于Meta基础设施芯片方向。Former Google TPU team lead, now working on infrastructure silicon at Meta.
A●● N●● TG-06TPU
TPU Chip Lead at Google
美国United StatesTPU Chip LeadGoogle
在职Google TPU芯片负责人,与TPU项目血统直接对应。Currently TPU chip lead at Google — a direct match to the TPU project lineage.

分布式存储与数据库系统工程师Distributed Storage & Database Systems

K●● A●● TG-07Storage/DB
Software Engineer @ IBM Ceph Team
英国United KingdomCephDistributed storage
IBM Ceph分布式存储团队的工程师。Distributed-storage engineer on IBM's Ceph team.
B●● H●● TG-08Storage/DB
Software Engineer | Distributed Systems | TiDB
美国United StatesTiDBDistributed Systems
TiDB分布式数据库方向的系统工程师。Systems engineer working on the TiDB distributed database.
s●● r●● TG-09Storage/DB
Principal Software Engineer, Database Internals and C++
加拿大CanadaDatabase internalsC++
数据库内核(database internals)研发的主任级工程师。Principal-level engineer working on database internals in C++.
R●● V●● TG-10Storage/DB
Staff Software Engineer working on Cloud Spanner at Google
美国United StatesCloud SpannerGoogle
在职Google Cloud Spanner团队,与Jeff Dean的Spanner项目血统直接对应。Currently on Google's Cloud Spanner team — a direct match to Jeff Dean's Spanner lineage.
M●● L●● TG-11Storage/DB
Software Engineer - Distributed Database at Huawei Canada
加拿大CanadaDistributed DatabaseHuawei Canada
华为加拿大的分布式数据库研发工程师。Distributed-database developer at Huawei Canada.
A●● G●● TG-12Storage/DB
Original creator of Apache DataFusion, Apache Arrow & DataFusion PMC Member
美国United StatesDataFusionApache Arrow PMC
Apache DataFusion的原创作者,是分布式数据库引擎领域的开源领导信号。Original creator of Apache DataFusion — an open-source leadership signal in distributed database engines.

分布式训练网络通信工程师(NCCL/RDMA)Training Networking (NCCL/RDMA)

S●● W●● A●● TG-13NCCL/RDMA
Cloud Network Engineer, HPC & AI Supercomputing, Azure, InfiniBand/RoCE
美国United StatesInfiniBand/RoCEAzure
Azure云上InfiniBand/RoCE高性能网络工程,是RPC系统血统在AI集群网络上的现代形态。InfiniBand/RoCE high-performance networking on Azure — the RPC-systems lineage reappearing as AI-cluster networking.
X●● L●● TG-14NCCL/RDMA
Senior HPC Performance Engineer, Networking - RDMA / GPU Communication - NCCL
美国United StatesNCCLGPU communication
专注NCCL/GPU通信的资深HPC性能工程师,与RPC系统血统直接对应。Senior HPC performance engineer focused on NCCL/GPU communication — a direct match to the RPC-systems lineage.
H●● M●● TG-15NCCL/RDMA
Senior Research Engineer, NTT Network Innovation Laboratories, RDMA/Infiniband Developer
日本JapanRDMAInfiniband
NTT网络创新实验室的RDMA/Infiniband研发工程师。RDMA/Infiniband developer at NTT's Network Innovation Laboratories.
R●● T●● TG-16NCCL/RDMA
RDMA Developer at China Telecom Corporation Limited
美国United StatesRDMAChina Telecom
中国电信的RDMA研发工程师。RDMA developer at China Telecom.
M●● G●● TG-17NCCL/RDMA
Lead Systems Engineer (HPC), NVIDIA InfiniBand RDMA Expert - NVL72 (NVLink)
美国United StatesNVLinkNVL72
NVIDIA InfiniBand RDMA专家,覆盖NVLink/NVL72最新一代GPU互联。NVIDIA InfiniBand RDMA expert covering the latest NVLink/NVL72 GPU-interconnect generation.
N●● M●● TG-18NCCL/RDMA
Senior Linux System Software Developer, Infiniband at SUSE
法国FranceInfinibandSUSE
SUSE的Infiniband系统软件资深研发工程师。Senior Infiniband systems-software developer at SUSE.

大语言模型研发工程师(预训练)LLM Pretraining Research

Q●● L●● TG-19LLM Pretraining
Senior Applied Scientist, Amazon AGI
美国United StatesApplied ScientistAmazon AGI
Amazon AGI的资深应用科学家。Senior applied scientist at Amazon AGI.
H●● Y●● TG-20LLM Pretraining
Research Scientist @ Meta, (Multimodal) LLM for Recommendation
美国United StatesMultimodal LLMMeta
Meta的多模态大语言模型研究科学家。Research scientist at Meta working on multimodal LLMs.
S●● V●● TG-21LLM Pretraining
RE @ GDM, RE @ Character AI, Senior MLE @ Square, AI Resident @ Facebook AI
美国United StatesGoogle DeepMindCharacter AI
历经Google DeepMind、Character AI等多家前沿大模型机构。Research-engineering experience across Google DeepMind, Character AI, and other frontier-model organizations.
Z●● B●● TG-22LLM Pretraining
Research Scientist @ Meta FAIR, LLM pretraining @ Amazon AGI/AWS
美国United StatesLLM pretrainingMeta FAIR
跨Meta FAIR和Amazon AGI两段大语言模型预训练研究经历。LLM pretraining research spanning both Meta FAIR and Amazon AGI/AWS.
T●● R●● TG-23LLM Pretraining
LLM pretraining research at the Allen Institute for AI
美国United StatesLLM pretrainingAI2
Allen Institute for AI的大语言模型预训练研究员。LLM pretraining researcher at the Allen Institute for AI.
H●● P●● TG-24LLM Pretraining
Research Scientist/Engineer II (Senior), Adobe Research, Foundation Model Pretraining
美国United StatesFoundation Model PretrainingAdobe Research
Adobe Research基础模型预训练方向的资深研究工程师。Senior research scientist/engineer at Adobe Research working on foundation-model pretraining.
Job-posting demand

03同类公司招聘正在补 Infrastructure / Platform 和 Domain SciencePeer-company hiring is filling Infrastructure / Platform and Domain Science

2026-08-08 的招聘数据包含 264 个招聘职位,招聘集中在 Infrastructure / Platform 和 Domain Science,资历段集中在 Mid-Senior level。需求侧偏向能搭建平台、承接实验、推进领域问题的成熟人才。As of 2026-08-08, there are 264 job postings, concentrated in Infrastructure / Platform and Domain Science, with seniority centered on Mid-Senior level. Demand skews toward mature talent that can build platforms, run experiments, and move domain problems forward.

招聘主题Hiring themes

Infrastructure / Platform 和 Domain Science 是前两类,Lab Automation 以 39 个岗位排在第三。平台、科学和实验自动化构成同类公司扩张时的主要招聘面。Infrastructure / Platform and Domain Science lead, with Lab Automation third at 39 postings — the main hiring surface for this peer set.

Infrastructure / Platform
96
Domain Science
85
Lab Automation
39
Research Engineering
25
AI / ML
9
Product / Ops / Business
5
Other / Unknown
5
数据来源: Metix AI(2026-08-08)Data source: Metix AI (2026-08-08)

招聘资历结构Seniority mix in hiring

Mid-Senior level 有 183 个岗位,Director 有 25 个,Entry level 有 16 个。招聘重心落在能直接承担系统、实验和项目推进的经验带。Mid-Senior level has 183 postings, Director has 25, and Entry level has 16. The hiring center of gravity sits in the experience band that can directly own systems, experiments, and project execution.

Mid-Senior level
183
Director
25
Entry level
16
Internship
13
Associate
12
Executive
9
Not Applicable
6
数据来源: Metix AI(2026-08-08)Data source: Metix AI (2026-08-08)
Company map

04公司地图显示人才记录集中在少数 AI for Science 公司The company map shows talent-record concentration in a few AI for Science companies

21 家同类公司覆盖五类方向,其中 AI for Science Platform 有 16 家。Recursion(758)、Lila Sciences(368)、Isomorphic Labs(361)、Generate:Biomedicines(325) 位列人才记录规模前四,构成第一层公司地图。The 21 peer companies span five lanes, with 16 in AI for Science Platform. Recursion (758), Lila Sciences (368), Isomorphic Labs (361), Generate:Biomedicines (325) rank as the top four talent-record pools and form the first layer of the company map.

同类公司人才池规模Peer-company talent-pool scale

21同类公司peer companies
3,612人才记录规模talent-record scale
264招聘职位job postings
AI for Science Platform16 家公司companies

模型、实验数据和科学问题位于同一条研发线。Models, experimental data, and scientific problems sit in one R&D line.

Autonomous Lab / Robotic Experiment Platform2 家公司companies

实验设备、自动化调度和结果回流是该类公司的主要能力信号。Lab equipment, automation scheduling, and result feedback are the main capability signals in this lane.

Foundation Model Research Infrastructure1 家公司companies

研究工程、训练平台、模型服务和评测平台是该类公司的主要能力信号。Research engineering, training platforms, model serving, and evaluation platforms are the main capability signals in this lane.

AI-designed Hardware / Materials Discovery1 家公司companies

材料、芯片、加速器和仿真方向构成该类公司的主要能力信号。Materials, chips, accelerators, and simulation form the main capability signals in this lane.

Automated AI Research Loop1 家公司companies

研究任务、评测和迭代机制产品化是该类公司的主要能力信号。Productized research tasks, evaluation, and iteration mechanisms are the main capability signals in this lane.

同类公司人才记录规模Talent-record scale in peer companies

公司地图将 automated discovery 的人才搜索压到少数高密度组织,减少宽泛 AI 公司池带来的噪音。The company map concentrates automated discovery talent search into a small set of high-density organizations and reduces noise from broad AI-company pools.

Recursion
758 United States
Lila Sciences
368 United States
Isomorphic Labs
361 United Kingdom
Generate:Biomedicines
325 United States
Insilico Medicine
285 United States
Xaira Therapeutics
202 United States
Absci
175 United States
Sakana AI
173 Japan
Genesis Therapeutics
155 United States
Iambic Therapeutics
144 United States
数据来源: Metix AIData source: Metix AI

Automated discovery 人才清单可继续下钻Automated discovery talent lists can go deeper

Metix AI 可按招聘关键词、公司、地区、资历和技术层级拆分人才记录,形成可核验的人才搜索清单。Metix AI can split talent records by hiring keyword, company, region, seniority, and technical layer to form a verifiable talent-search list.

可继续展开公司、岗位、技术栈和候选人匹配。Company, job, stack, and candidate-fit views can be expanded.